
# Librerías básicas
import numpy as np
import pandas as pd
import seaborn as sns
from sklearn import datasets
from sklearn import preprocessing
from sklearn.manifold import TSNE
from sklearn.decomposition import PCA
from sklearn.ensemble import RandomForestRegressor, RandomForestClassifier
from sklearn.metrics import mean_absolute_error, mean_squared_error, mean_absolute_percentage_error, make_scorer, accuracy_score, roc_auc_score
from sklearn.model_selection import train_test_split, cross_val_score
from sklearn.model_selection import validation_curve
from sklearn.model_selection import GridSearchCV

import matplotlib.pyplot as plt

pd.set_option('display.max_columns', None)
seed = 100

%matplotlib inline


diabetes = datasets.load_diabetes()

print("Elements of diabetes: ", diabetes.keys())

Elements of diabetes:  dict_keys(['data', 'target', 'frame', 'DESCR', 'feature_names', 'data_filename', 'target_filename', 'data_module'])


print(diabetes["DESCR"])

.. _diabetes_dataset:

Diabetes dataset
----------------

Ten baseline variables, age, sex, body mass index, average blood
pressure, and six blood serum measurements were obtained for each of n =
442 diabetes patients, as well as the response of interest, a
quantitative measure of disease progression one year after baseline.

**Data Set Characteristics:**

  :Number of Instances: 442

  :Number of Attributes: First 10 columns are numeric predictive values

  :Target: Column 11 is a quantitative measure of disease progression one year after baseline

  :Attribute Information:
      - age     age in years
      - sex
      - bmi     body mass index
      - bp      average blood pressure
      - s1      tc, total serum cholesterol
      - s2      ldl, low-density lipoproteins
      - s3      hdl, high-density lipoproteins
      - s4      tch, total cholesterol / HDL
      - s5      ltg, possibly log of serum triglycerides level
      - s6      glu, blood sugar level

Note: Each of these 10 feature variables have been mean centered and scaled by the standard deviation times `n_samples` (i.e. the sum of squares of each column totals 1).

Source URL:
https://www4.stat.ncsu.edu/~boos/var.select/diabetes.html

For more information see:
Bradley Efron, Trevor Hastie, Iain Johnstone and Robert Tibshirani (2004) "Least Angle Regression," Annals of Statistics (with discussion), 407-499.
(https://web.stanford.edu/~hastie/Papers/LARS/LeastAngle_2002.pdf)


# desem els atributs i el target a les variables X i y

X = diabetes.data
y = diabetes.target

# Número de files

print("Number of rows X: ", X.shape[0])
print("Number of rows Y: ", y.shape[0])

# Número de columnes

print("Number of columns X: ", X.shape[1])

# Nombre de les columnes

l_feat_names = diabetes.feature_names
print("Columns with names: ", len(l_feat_names))
print("Columns names: ", l_feat_names)

Number of rows X:  442
Number of rows Y:  442
Number of columns X:  10
Columns with names:  10
Columns names:  ['age', 'sex', 'bmi', 'bp', 's1', 's2', 's3', 's4', 's5', 's6']


# Creem un dataframe amb les dades

df_diabetes = pd.concat([pd.DataFrame(X, columns=l_feat_names), 
                         pd.DataFrame(y, columns=["target"])], axis=1)

print("Number of rows: ", df_diabetes.shape[0], " and number of columns: ", df_diabetes.shape[1])
df_diabetes.head()

Number of rows:  442  and number of columns:  11


# Cerquem valors nuls o perduts

df_diabetes.isnull().sum()

age       0
sex       0
bmi       0
bp        0
s1        0
s2        0
s3        0
s4        0
s5        0
s6        0
target    0
dtype: int64


df_diabetes.dtypes

age       float64
sex       float64
bmi       float64
bp        float64
s1        float64
s2        float64
s3        float64
s4        float64
s5        float64
s6        float64
target    float64
dtype: object


df_diabetes.loc[df_diabetes["sex"] < 0, "sex"] = 0
df_diabetes.loc[df_diabetes["sex"] > 0, "sex"] = 1

freq = df_diabetes["sex"].value_counts()
freq

0.0    235
1.0    207
Name: sex, dtype: int64


plt.bar(freq.index.values, freq, align='center', alpha=0.6, color=['royalblue'])

plt.xticks(freq.index.values, ('Male ('+str(df_diabetes["sex"].value_counts()[0])+')',
                   'Female ('+str(df_diabetes["sex"].value_counts()[1])+')'))

plt.ylabel('Counts')
plt.title('Sex')

Text(0.5, 1.0, 'Sex')


df_diabetes.drop(columns=['sex', 'target']).describe()


df_diabetes.target.describe()

count    442.000000
mean     152.133484
std       77.093005
min       25.000000
25%       87.000000
50%      140.500000
75%      211.500000
max      346.000000
Name: target, dtype: float64


# Histograma
fig, ax = plt.subplots(nrows=3, ncols=3, figsize=(20, 10))

# Iterar con un solo indice
ax = ax.flatten()

for i, feature in enumerate(df_diabetes.drop(columns=['sex', 'target']).columns):
    ax[i].hist(df_diabetes[feature], alpha=0.6, color='royalblue', bins=10)
    ax[i].set_title(feature.upper())


response = "target"
cat_feats = "sex"
num_feats = ['age', 'bmi', 'bp', 's1', 's2', 's3', 's4', 's5', 's6']
print('Respuesta (target)    :', response)
print("Atributos categóricos :", cat_feats)
print("Atributos numéricos   :", num_feats)

Respuesta (target)    : target
Atributos categóricos : sex
Atributos numéricos   : ['age', 'bmi', 'bp', 's1', 's2', 's3', 's4', 's5', 's6']


df_diabetes.corr()['target'].sort_values(ascending=False)

target    1.000000
bmi       0.586450
s5        0.565883
bp        0.441484
s4        0.430453
s6        0.382483
s1        0.212022
age       0.187889
s2        0.174054
sex       0.043062
s3       -0.394789
Name: target, dtype: float64


corr_all = df_diabetes[num_feats + [response]].corr()
corr_all.style.background_gradient(cmap='coolwarm').set_precision(2)

C:\Users\oriol\AppData\Local\Temp\ipykernel_17688\4234408769.py:2: FutureWarning: this method is deprecated in favour of `Styler.format(precision=..)`
  corr_all.style.background_gradient(cmap='coolwarm').set_precision(2)


sns.pairplot(df_diabetes, 
            diag_kind= "kde", 
            hue= ('sex'),
            vars= num_feats
            )

<seaborn.axisgrid.PairGrid at 0x1d9c0ad00d0>


corr = df_diabetes.corr()
corr[['target']].sort_values(by=['target'], ascending=False).style.background_gradient()


sns.set_theme(style="white", palette="pastel")
fig, axs = plt.subplots(nrows = 2, ncols = 2, figsize = (16, 10))
axs = axs.flatten()
corr_vars = ['bmi', 's5', 'sex', 's2']

for i, var in enumerate(corr_vars):
    sns.regplot(x = var,
                y = 'target', 
                data = df_diabetes, 
                ax = axs[i],
                color='royalblue')
    
    #axs[i].set_title("Correlación con la variable objetivo", fontsize = 15)
    axs[i].set_xlabel(var, fontsize = 12)
    axs[i].set_ylabel('target', fontsize = 12)
plt.tight_layout()


X = df_diabetes.drop(columns =[response])

y = df_diabetes[response].values


from sklearn.preprocessing import StandardScaler

scaler = StandardScaler()
X_scaled = scaler.fit_transform(X)


pca = PCA(n_components=2, random_state=seed)
X_pca = pca.fit_transform(X_scaled)

print("Dimensions de les dades reduïdes amb PCA:", np.shape(X_pca))
print('Explained variance ratio (first two components): %s' % str(pca.explained_variance_ratio_))

Dimensions de les dades reduïdes amb PCA: (442, 2)
Explained variance ratio (first two components): [0.40242142 0.14923182]


fig, ax = plt.subplots(nrows=1, ncols=1, figsize=(8,6))

scatter = ax.scatter(x=X_pca[:, 0], y=X_pca[:, 1], c=y, cmap='viridis')
ax.set_xlabel('1st principal component')
ax.set_ylabel('2nd principal component')
ax.set_title('PCA of diabetes dataset')

colorbar = plt.colorbar(scatter)

plt.tight_layout()


tsne = TSNE(n_components=2, random_state=seed, perplexity=40, init='random', learning_rate=150)
X_tsne = tsne.fit_transform(X_scaled)


fig, ax = plt.subplots(nrows=1, ncols=1, figsize=(8,6))

scatter = ax.scatter(x=X_tsne[:, 0], y=X_tsne[:, 1], c=y, cmap='viridis')
ax.set_xlabel('1st principal component')
ax.set_ylabel('2nd principal component')
ax.set_title('TSNE of diabetes dataset')

colorbar = plt.colorbar(scatter)

plt.tight_layout()


clf = RandomForestRegressor(n_estimators=10, random_state=seed)

scores = cross_val_score(clf, X, y, cv=5, scoring=make_scorer(mean_absolute_error))
mean_data = np.abs(np.mean(scores))
std_data = np.std(scores)
print('mean MAE score: {:.1f}, std: {:.1f}, 1std score range: {:.1f} - {:.1f}'.format(mean_data, std_data, mean_data - std_data, mean_data + std_data))

scores = cross_val_score(clf, X, y, cv=5, scoring=make_scorer(mean_squared_error))
mean_data = np.abs(np.mean(scores))
std_data = np.std(scores)
print('mean MSE score: {:.1f}, std: {:.1f}, 1std score range: {:.1f} - {:.1f}'.format(mean_data, std_data, mean_data - std_data, mean_data + std_data))

scores = cross_val_score(clf, X, y, cv=5, scoring=make_scorer(mean_absolute_percentage_error))
mean_data = np.abs(np.mean(scores))
std_data = np.std(scores)
print('mean MAPE score: {:.1f}, std: {:.1f}, 1std score range: {:.1f} - {:.1f}'.format(mean_data, std_data, mean_data - std_data, mean_data + std_data))

mean MAE score: 48.4, std: 2.1, 1std score range: 46.3 - 50.5
mean MSE score: 3640.9, std: 289.5, 1std score range: 3351.4 - 3930.4
mean MAPE score: 0.4, std: 0.0, 1std score range: 0.4 - 0.5


X = df_diabetes.drop(columns=['target', 'sex']).values
y = df_diabetes['sex'].astype(int)


clf = RandomForestClassifier(n_estimators=10)

scores = cross_val_score(clf, X, y, cv=5, scoring=make_scorer(accuracy_score))
mean_data = np.abs(np.mean(scores))
std_data = np.std(scores)
print('Accuracy score: {:.1f}, std: {:.1f}, 1std score range: {:.1f} - {:.1f}'.format(mean_data, std_data, mean_data - std_data, mean_data + std_data))

scores = cross_val_score(clf, X, y, cv=5, scoring=make_scorer(roc_auc_score))
mean_data = np.abs(np.mean(scores))
std_data = np.std(scores)
print('ROC AUC score: {:.1f}, std: {:.1f}, 1std score range: {:.1f} - {:.1f}'.format(mean_data, std_data, mean_data - std_data, mean_data + std_data))

Accuracy score: 0.7, std: 0.0, 1std score range: 0.6 - 0.7
ROC AUC score: 0.7, std: 0.0, 1std score range: 0.6 - 0.7

	age	sex	bmi	bp	s1	s2	s3	s4	s5	s6	target
0	0.038076	0.050680	0.061696	0.021872	-0.044223	-0.034821	-0.043401	-0.002592	0.019908	-0.017646	151.0
1	-0.001882	-0.044642	-0.051474	-0.026328	-0.008449	-0.019163	0.074412	-0.039493	-0.068330	-0.092204	75.0
2	0.085299	0.050680	0.044451	-0.005671	-0.045599	-0.034194	-0.032356	-0.002592	0.002864	-0.025930	141.0
3	-0.089063	-0.044642	-0.011595	-0.036656	0.012191	0.024991	-0.036038	0.034309	0.022692	-0.009362	206.0
4	0.005383	-0.044642	-0.036385	0.021872	0.003935	0.015596	0.008142	-0.002592	-0.031991	-0.046641	135.0

	age	bmi	bp	s1	s2	s3	s4	s5	s6
count	4.420000e+02	4.420000e+02	4.420000e+02	4.420000e+02	4.420000e+02	4.420000e+02	4.420000e+02	4.420000e+02	4.420000e+02
mean	-3.634285e-16	-8.045349e-16	1.281655e-16	-8.835316e-17	1.327024e-16	-4.574646e-16	3.777301e-16	-3.830854e-16	-3.412882e-16
std	4.761905e-02	4.761905e-02	4.761905e-02	4.761905e-02	4.761905e-02	4.761905e-02	4.761905e-02	4.761905e-02	4.761905e-02
min	-1.072256e-01	-9.027530e-02	-1.123996e-01	-1.267807e-01	-1.156131e-01	-1.023071e-01	-7.639450e-02	-1.260974e-01	-1.377672e-01
25%	-3.729927e-02	-3.422907e-02	-3.665645e-02	-3.424784e-02	-3.035840e-02	-3.511716e-02	-3.949338e-02	-3.324879e-02	-3.317903e-02
50%	5.383060e-03	-7.283766e-03	-5.670611e-03	-4.320866e-03	-3.819065e-03	-6.584468e-03	-2.592262e-03	-1.947634e-03	-1.077698e-03
75%	3.807591e-02	3.124802e-02	3.564384e-02	2.835801e-02	2.984439e-02	2.931150e-02	3.430886e-02	3.243323e-02	2.791705e-02
max	1.107267e-01	1.705552e-01	1.320442e-01	1.539137e-01	1.987880e-01	1.811791e-01	1.852344e-01	1.335990e-01	1.356118e-01

	age	bmi	bp	s1	s2	s3	s4	s5	s6	target
age	1.00	0.19	0.34	0.26	0.22	-0.08	0.20	0.27	0.30	0.19
bmi	0.19	1.00	0.40	0.25	0.26	-0.37	0.41	0.45	0.39	0.59
bp	0.34	0.40	1.00	0.24	0.19	-0.18	0.26	0.39	0.39	0.44
s1	0.26	0.25	0.24	1.00	0.90	0.05	0.54	0.52	0.33	0.21
s2	0.22	0.26	0.19	0.90	1.00	-0.20	0.66	0.32	0.29	0.17
s3	-0.08	-0.37	-0.18	0.05	-0.20	1.00	-0.74	-0.40	-0.27	-0.39
s4	0.20	0.41	0.26	0.54	0.66	-0.74	1.00	0.62	0.42	0.43
s5	0.27	0.45	0.39	0.52	0.32	-0.40	0.62	1.00	0.46	0.57
s6	0.30	0.39	0.39	0.33	0.29	-0.27	0.42	0.46	1.00	0.38
target	0.19	0.59	0.44	0.21	0.17	-0.39	0.43	0.57	0.38	1.00

PEC 1: Preparación de datos¶

0. Carga de librerías¶

1. Carga del conjunto de datos (1 punto)¶

2. Análisis de los datos (3 puntos)¶

2.1 Análisis estadístico básico¶

2.2 Análisis exploratorio de los datos¶

3. Reducción de la dimensionalidad (3 puntos)¶

4 Entrenamiento y métricas (3 puntos)¶

4.a Regresión¶

4.b Clasificación¶

	target
target	1.000000
bmi	0.586450
s5	0.565883
bp	0.441484
s4	0.430453
s6	0.382483
s1	0.212022
age	0.187889
s2	0.174054
sex	0.043062
s3	-0.394789